Semantic Frame Identification with Distributed Word Representations

By Karl Moritz Hermann, Dipanjan Das, Jason Aaron Edward Weston, Kuzman Ganchev. 2014

128 citations recorded by the source corpus.

We present a novel technique for semantic frame identiﬁcation using distributed representations of predicates and their syntactic context; this technique leverages automatic syntactic parses and a generic set of word embeddings. Given labeled data annotated with frame-semantic parses, we learn a model that projects the set of word representations for the syntactic context around a predicate to a low dimensional representation. The latter is used for semantic frame identiﬁcation; with a standard argument identiﬁcation method inspired by prior work, we achieve state-of-the-art results on FrameNet-style frame-semantic analysis. Additionally, we report strong results on PropBank-style semantic role labeling in comparison to prior work.

Subfields: Artificial Intelligence, Information Extraction

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Information Extraction

Citations in the corpus

StarSpace: Embed All The Things! (2017, background): “Inthedomain of supervised embeddings, SSI (Bai et al. 2009) and WSABIE (Weston, Bengio, and Usunier 2011) are early approaches that showed promise in NLP and information retrieval tasks ((Weston et al. 2013), (Hermann et al. 2014)).”
Factor-based Compositional Embedding Models (2014, background): “They have been successfully applied to a range of NLP tasks, including syntax [2, 4, 5] and semantics [6, 7, 8].”
Efficient Inference and Structured Learning for Semantic Role Labeling (2015, methodology): “For FrameNet, we use the extraction method of Hermann et al. (2014, §5.4), which is also inspired by Xue and Palmer (2004).”
Multitask learning for semantic sequence prediction under varying data conditions (2016, unknown): “We use the data splits from (Das et al., 2014; Hermann et al., 2014).”
FrameNet+: Fast Paraphrastic Tripling of FrameNet (2015, background): “Das and Smith (2011) and Das and Smith (2012) used graph based semi-supervised methods to improve frame coverage and Hermann et al. (2014) used word and frame embeddings to improve generalization.”
User Conditional Hashtag Prediction for Images (2015, methodology): “Of particular relevance to our work is Wsabie, a supervised embedding approach that has been used for large scale image annotation and NLP tasks [17, 3].”
A Deep Architecture for Semantic Parsing (2014, background): “, 2014; Hermann and Blunsom, 2014a), frame-semantic parsing (Hermann et al., 2014), and machine translation (Mikolov et al.”
A Joint Sequential and Relational Model for Frame-Semantic Parsing (2017, methodology): “For baselines, we consider the prior state-of-the-art approach WSABIE EMBEDDING (Hermann et al., 2014), which learns feature representations based on word embeddings and dependency path embeddings using the WSABIE algorithm (Weston et al., 2011).”
#TagSpace: Semantic Embeddings from Hashtags (2014, background): “WSABIE (Weston et al., 2011) is a supervised embedding approach that has shown promise in NLP tasks (Weston et al., 2013; Hermann et al., 2014).”
Frame-Semantic Parsing (2014, background): “Since Gildea and Jurafsky (2002) pioneered statistical semantic role labeling, there has been a great deal of computational work using predicate-argument structures for semantics.”